Searcharxiv⌕ Search

arXiv subjects

Tsuyoshi Kato

Publications and source records attributed to Tsuyoshi Kato.

At least 19 recordsLinked to original sources
arXiv2026

Moduli Spaces of Connections and B-fields from T-duality with H-flux

We study the geometry of mixed fields, consisting of a connection and a $B$-field on a principal circle bundle over a Riemann surface, from the perspective of gauge theory and T-duality. Motivated by the foundational work of Atiyah--Bott and Segal, we introduce a twisted Yang--Mills functional whose critical locus, in the flat case, is governed by the simultaneous vanishing of the curvature and the $H$-flux. We show that the gauge group is a semi-direct product of abelian groups parametrised by an integer $\lambda$. The moduli spaces are constructed by presymplectic reduction and shown to be Heisenberg contact manifolds for $\lambda \neq 0$, whose topology we characterise completely. We show that T-duality preserves the twisted Yang--Mills functional and acts on the configuration space of flat mixed fields. We identify the subgroups of gauge transformations that are compatible with the T-duality map and describe the induced action on the singular quotient of T-dualizable flat mixed fields. Precisely at $\lambda =1$ does T-duality descend to an involutive contactomorphism of the moduli space.

hep-th↗
arXiv2025

Notions of simple type for Bauer--Furuta invariants

By extending the notion of simple type for the Seiberg--Witten invariant of a 4-manifold, we introduce notions of BF blowup simple type and BF homogeneous type for the Bauer--Furuta invariant and study their applications. Specifically, we show that the existence of an immersed 2-sphere with a certain condition guarantees BF blowup simple type. As an application, we determine the Bauer--Furuta invariant of a 4-manifold obtained by a logarithmic transformation along a torus in a fishtail neighborhood. We also give constraints on gluing decompositions of 4-manifolds by using BF homogeneous type. To prove these results, we also give gluing formulae and an immersed adjunction inequality for Bauer--Furuta invariants.

math.GT↗
arXiv2025

Uniform Lefschetz fixed-point theory

We develop the Lefschetz fixed-point theory for noncompact manifolds of bounded geometry and uniformly continuous maps. Specifically, we define the uniform Lefschetz class $\mathscr{L}(f)$ of a uniformly continuous map $f\colon M\to M$ of a uniform simply-connected noncompact complete Riemannian manifold of bounded geometry $M$ satisfying $d(f,1)<\infty$, and prove that $\mathscr{L}(f)=0$ if and only if $f$ is uniformly homotopic to a strongly fixed-point free (without fixed-points on $M$ and at infinity) uniformly continuous map. To achieve this, we introduce a new cohomology for metric spaces, called uniform bounded cohomology, which is a variant of bounded cohomology, and develop an obstruction theory formulated in terms of this cohomology.

math.AT↗
arXiv2025

Morse inequalities for noncompact manifolds

We establish Morse inequalities for a noncompact manifold with a cocompact and properly discontinuous action of a discrete group, where Morse functions are not necessarily invariant under the group action. The inequalities are given in terms of the $L^2$-Betti numbers and functions on the acting group which describe rough configurations of critical points of a Morse function.

math.GT↗
arXiv2025

Linearly Convergent Mixup Learning

Learning in the reproducing kernel Hilbert space (RKHS) such as the support vector machine has been recognized as a promising technique. It continues to be highly effective and competitive in numerous prediction tasks, particularly in settings where there is a shortage of training data or computational limitations exist. These methods are especially valued for their ability to work with small datasets and their interpretability. To address the issue of limited training data, mixup data augmentation, widely used in deep learning, has remained challenging to apply to learning in RKHS due to the generation of intermediate class labels. Although gradient descent methods handle these labels effectively, dual optimization approaches are typically not directly applicable. In this study, we present two novel algorithms that extend to a broader range of binary classification models. Unlike gradient-based approaches, our algorithms do not require hyperparameters like learning rates, simplifying their implementation and optimization. Both the number of iterations to converge and the computational cost per iteration scale linearly with respect to the dataset size. The numerical experiments demonstrate that our algorithms achieve faster convergence to the optimal solution compared to gradient descent approaches, and that mixup data augmentation consistently improves the predictive performance across various loss functions.

cs.LG↗
arXiv2024

A Study on Unsupervised Anomaly Detection and Defect Localization using Generative Model in Ultrasonic Non-Destructive Testing

In recent years, the deterioration of artificial materials used in structures has become a serious social issue, increasing the importance of inspections. Non-destructive testing is gaining increased demand due to its capability to inspect for defects and deterioration in structures while preserving their functionality. Among these, Laser Ultrasonic Visualization Testing (LUVT) stands out because it allows the visualization of ultrasonic propagation. This makes it visually straightforward to detect defects, thereby enhancing inspection efficiency. With the increasing number of the deterioration structures, challenges such as a shortage of inspectors and increased workload in non-destructive testing have become more apparent. Efforts to address these challenges include exploring automated inspection using machine learning. However, the lack of anomalous data with defects poses a barrier to improving the accuracy of automated inspection through machine learning. Therefore, in this study, we propose a method for automated LUVT inspection using an anomaly detection approach with a diffusion model that can be trained solely on negative examples (defect-free data). We experimentally confirmed that our proposed method improves defect detection and localization compared to general object detection algorithms used previously.

cs.CV↗
arXiv2023

Vector fields on non-compact manifolds

Let $M$ be a non-compact connected manifold with a cocompact and properly discontinuous action of a discrete group $G$. We establish a Poincaré-Hopf theorem for a bounded vector field on $M$ satisfying a mild condition on zeros. As an application, we show that such a vector field must have infinitely many zeros whenever $G$ is amenable and the Euler characteristic of $M/G$ is non-zero.

math.GT↗
arXiv2023

Homotopy type of the space of finite propagation unitary operators on $\mathbb{Z}$

The index theory for the space of finite propagation unitary operators was developed by Gross, Nesme, Vogts and Werner from the viewpoint of quantum walks in mathematical physics. In particular, they proved that $π_0$ of the space is determined by the index. However, nothing is known about the higher homotopy groups. In this article, we describe the homotopy type of the space of finite propagation unitary operators on the Hilbert space of square summable $\mathbb{C}$-valued $\mathbb{Z}$-sequences, so we can determine its homotopy groups. We also study the space of (end-)periodic finite propagation unitary operators.

math.AT↗
arXiv2023

Simulation-Aided Deep Learning for Laser Ultrasonic Visualization Testing

In recent years, laser ultrasonic visualization testing (LUVT) has attracted much attention because of its ability to efficiently perform non-contact ultrasonic non-destructive testing.Despite many success reports of deep learning based image analysis for widespread areas, attempts to apply deep learning to defect detection in LUVT images face the difficulty of preparing a large dataset of LUVT images that is too expensive to scale. To compensate for the scarcity of such training data, we propose a data augmentation method that generates artificial LUVT images by simulation and applies a style transfer to simulated LUVT images.The experimental results showed that the effectiveness of data augmentation based on the style-transformed simulated images improved the prediction performance of defects, rather than directly using the raw simulated images for data augmentation.

eess.IV↗
arXiv2023

A Study on Deep CNN Structures for Defect Detection From Laser Ultrasonic Visualization Testing Images

The importance of ultrasonic nondestructive testing has been increasing in recent years, and there are high expectations for the potential of laser ultrasonic visualization testing, which combines laser ultrasonic testing with scattered wave visualization technology. Even if scattered waves are visualized, inspectors still need to carefully inspect the images. To automate this, this paper proposes a deep neural network for automatic defect detection and localization in LUVT images. To explore the structure of a neural network suitable to this task, we compared the LUVT image analysis problem with the generic object detection problem. Numerical experiments using real-world data from a SUS304 flat plate showed that the proposed method is more effective than the general object detection model in terms of prediction performance. We also show that the computational time required for prediction is faster than that of the general object detection model.

cs.CV↗
arXiv2023

Upper bounds for virtual dimensions of Seiberg-Witten moduli spaces

Given a closed four-manifold with $b_1=0$ and a prime number $p$, we prove that for any mod $p^r$ basic class, the virtual dimension of the Seiberg-Witten moduli space is bounded above by $2r(p-1)-2$ under some conditions on $r$ and $b_2^+$. As an application, we obtain adjunction inequalities for embedded surfaces with negative self-intersection number.

math.GT↗
arXiv2023

Multi-Target Tobit Models for Completing Water Quality Data

Monitoring microbiological behaviors in water is crucial to manage public health risk from waterborne pathogens, although quantifying the concentrations of microbiological organisms in water is still challenging because concentrations of many pathogens in water samples may often be below the quantification limit, producing censoring data. To enable statistical analysis based on quantitative values, the true values of non-detected measurements are required to be estimated with high precision. Tobit model is a well-known linear regression model for analyzing censored data. One drawback of the Tobit model is that only the target variable is allowed to be censored. In this study, we devised a novel extension of the classical Tobit model, called the \emph{multi-target Tobit model}, to handle multiple censored variables simultaneously by introducing multiple target variables. For fitting the new model, a numerical stable optimization algorithm was developed based on elaborate theories. Experiments conducted using several real-world water quality datasets provided an evidence that estimating multiple columns jointly gains a great advantage over estimating them separately.

cs.AI↗
arXiv2022

Seiberg-Witten theory on finite covering spaces of spin 4-manifolds

We compute the equivariant Bauer-Furuta degree, when a finite group acts freely on a spin 4-manifold. In the case when the group is cyclic of order power of two, Bryan gave a formula and its applications. We have treated the case when the group has order of odd degree. In particular we gave a formula of the degree when the order is odd-prime. Our approach is to use a representation-theoretic method on finite dimensional approximations of the functional spaces.

math.KT↗
arXiv2021

A note on exotic families of 4-manifolds

We present a pair of smooth fiber bundles over the circle with a common $4$-dimensional fiber with the following properties: (1) their total spaces are diffeomorphic to each other; (2) they are isomorphic to each other as topological fiber bundles; (3) they are not isomorphic to each other as smooth fiber bundles. In particular, we exhibit an example with non-simply-connected fiber.

math.GT↗
arXiv2021

$L^2$ harmonic theory, Seiberg-Witten theory and asymptotics of differential forms

We present a pair of open smooth $4$-manifolds that are mutually homeomorphic. One of them admits a Riemannian metric that possesses quasi-cylindricity, and positivity of scalar curvature and of dimension of certain $L^2$ harmonic forms. By contrast, for the other manifold, no Riemannian metric can simultaneously satisfy these properties. Our method uses Seiberg-Witten theory on compact $4$-manifolds and applies $L^2$ harmonic theory on non-compact, complete Riemannian $4$-manifolds. We introduce a new argument to apply Gauge theory, which arises from a discovery of an asymptotic property of the range of the differential.

math.DG↗
arXiv2021

Hilbert bundles with ends

Given a countable metric space, we can consider its end. Then a basis of a Hilbert space indexed by the metric space defines an end of the Hilbert space, which is a new notion and different from an end as a metric space. Such an indexed basis also defines unitary operators of finite propagation, and these operators preserve an end of a Hilbert space. Then, we can define a Hilbert bundle with end, which lightens up new structures of Hilbert bundles. In a special case, we can define characteristic classes of Hilbert bundles with ends, which are new invariants of Hilbert bundles. We show Hilbert bundles with ends appear in natural contexts. First, we generalize the pushforward of a vector bundle along a finite covering to an infinite covering, which is a Hilbert bundle with end under a mild condition. Then we compute characteristic classes of some pushforwards along infinite coverings. Next, we will show the spectral decompositions of nice differential operators give rise to Hilbert bundles with ends, which elucidate new features of spectral decompositions. The spectral decompositions we will consider are the Fourier transform and the harmonic oscillators.

math.KT↗
arXiv2021

Covering monopole map and higher degree in non commutative geometry

In this paper we develop analysis of the monopole maps over the universal covering space of a compact four manifold. We induce a property on local properness of the covering monopole map under the condition of closeness of the AHS complex. In particular we construct a higher degree of the covering monopole map when the linearized equation is isomorphic, which induces a homomorphism between K group of the group C^* algebras. It involves non linear analysis on the covering space, which is related to L^p cohomology. We also obtain various Sobolev estimates on the covering spaces. As a possible application, we propose an aspherical version of 10/8 inequality, combining with Singer conjecture on L^2 cohomology. It is satisfied for a large class of four manifolds which includes some complex surfaces of general type.

math.DG↗